Exact Asymptotics for a Multitimescale Model with Applications in Modeling Overdispersed Customer Streams

Large deviations (60F10) Queues and service in operations research (90B22) Queueing theory (aspects of probability theory) (60K25) Performance evaluation, queueing, and scheduling in the context of computer systems (68M20)

Abstract: In this paper we study the probability

x i_{n} (u) := m a t h b b P l e f t (C_{n} g e q s l a n t u n i g h t)

, with

C_{n} := A (p s i_{n} B (v a r p h i_{n}))

for L'{e}vy processes

A (c d o t)

and

B (c d o t)

, and

v a r p h i_{n}

and

p s i_{n}

non-negative sequences such that

v a r p h i_{n} p s i_{n} = n

and

v a r p h i_{n} o i n f t y

as

n o i n f t y

. Two timescale regimes are distinguished: a `fast' regime in which

v a r p h i_{n}

is superlinear and a `slow' regime in which

v a r p h i_{n}

is sublinear. We provide the exact asymptotics of

x i_{n} (u)

(as

n o i n f t y

) for both regimes, relying on change-of-measure arguments in combination with Edgeworth-type estimates. The asymptotics have an unconventional form: the exponent contains the commonly observed linear term, but may also contain sublinear terms (the number of which depends on the precise form of

v a r p h i_{n}

and

p s i_{n}

). To showcase the power of our results we include two examples, covering both the case where

C_{n}

is lattice and non-lattice. Finally we present numerical experiments that demonstrate the importance of taking into account the doubly stochastic nature of

C_{n}

in a practical application related to customer streams in service systems; they show that the asymptotic results obtained yield highly accurate approximations, also in scenarios in which there is no pronounced timescale separation.

Recommendations

Cites work

Cited in

(1)

Exact asymptotics in an infinite-server system with overdispersed input

This page was built for publication: Exact Asymptotics for a Multitimescale Model with Applications in Modeling Overdispersed Customer Streams

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5113898)